ary pharmacology &
TRANSCRIPT
This is a repository copy of Studies of salivary pepsin in patients with gastro oesophageal ‐
reflux disease.
White Rose Research Online URL for this paper:http://eprints.whiterose.ac.uk/140278/
Version: Accepted Version
Article:
Race, C., Chowdry, J., Russell, J. et al. (2 more authors) (2019) Studies of salivary pepsin in patients with gastro oesophageal reflux disease. Alimentary Pharmacology & ‐
Therapeutics, 49 (9). pp. 1173-1180. ISSN 0269-2813
https://doi.org/10.1111/apt.15138
This is the peer reviewed version of the following article: Race, C, Chowdry, J, Russell, JM,Corfe, BM, Riley, SA. Studies of salivary pepsin in patients with gastro oesophageal reflux ‐
disease. Aliment Pharmacol Ther. 2019; 49: 1173– 1180, which has been published in finalform at: https://doi.org/10.1111/apt.15138. This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Use of Self-Archived Versions.
[email protected]://eprints.whiterose.ac.uk/
Reuse
Items deposited in White Rose Research Online are protected by copyright, with all rights reserved unless indicated otherwise. They may be downloaded and/or printed for private study, or other acts as permitted by national copyright laws. The publisher or other rights holders may allow further reproduction and re-use of the full text version. This is indicated by the licence information on the White Rose Research Online record for the item.
Takedown
If you consider content in White Rose Research Online to be in breach of UK law, please notify us by emailing [email protected] including the URL of the record and the reason for the withdrawal request.
For Peer Review
Studies of Salivary Pepsin in Patients with GORD
Journal: Alimentary Pharmacology & Therapeutics
Manuscript ID APT-1523-2018.R1
Wiley - Manuscript type: Original Scientific Paper
Date Submitted by the
Author:06-Dec-2018
Complete List of Authors: Riley, Stuart; Sheffield, Gastroenterology
Race, Caroline ; Sheffield Teaching Hospitals NHS Foundation Trust,
Sheffield, Department of Gastrointestinal Physiology
Chowdry, Joanna; University of Sheffield, Molecular Gastroenterology
Research Group, Academic Unit of Surgical Oncology, Department of
Oncology and Metabolism
Russell, Jean; University of Sheffield, Department of Corporate
Information and Computer Services
Corfe, Bernard; University of Sheffield, Oncology and Metabolism
Keywords:GERD or GORD < Disease-based, Oesophagus < Organ-based,
Diagnostic tests < Topics, General practice < Topics
Alimentary Pharmacology & Therapeutic
For Peer Review
Studies of Salivary Pepsin in Patients with GERD
Short title: Salivary Pepsin in GERD
Caroline Race1,5 Joanna Chowdry2,5 Jean M. Russell3, Bernard M.
Corfe2,4,6 & Stuart A. Riley1,4,6,7
1. Department of Gastroenterology, Northern General Hospital, Herries Road, Sheffield
2. Molecular Gastroenterology Research Group, Academic Unit of Surgical Oncology,
Department of Oncology and Metabolism, University of Sheffield
3. Department of Corporate Information and Computer Services, University of Sheffield,
Hownslow Road, Sheffield
4. Insigneo Institute, University of Sheffield
5. These authors made an equal contribution to this work
6. These authors made an equal contribution to this work
7. Author for correspondence: Dr Stuart Riley
Keywords:
Word count: 3,006 words
Page 1 of 24 Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
SUMMARY
Background
Gastro-oesophageal reflux disease (GERD) is difficult to diagnose without invasive testing.
Peptest is a newly marketed diagnostic tool which aims to quantify salivary pepsin as a
marker of reflux, providing a rapid alternative to invasive procedures. We aimed to evaluate
optimal timing for sampling and to evaluate the accuracy of Peptest against an independent
measure.
Methods
Thirty diagnosed GERD patients (12 female, mean age 49 (range 20-72)) and twenty
asymptomatic subjects (14 female, mean age 56 (range 21-56)) were subject to diurnal
saliva sampling, with additional samples for sixty minutes following self-reported reflux
symptoms and following triggering of a proximal reflux alarm. Saliva samples were split and
were analysed by both Peptest and ELISA with operators for each blinded to sample identity.
Results
Salivary pepsin was detectable in most patients and most volunteers. Peptest scores were
significantly lower for patients than controls (P<0.005). ELISA scores showed no difference
between patients and controls. There was no effect of diurnal sampling time (P=0.75) or time
after symptoms (P=0.76) on Peptest readout. There was no correlation between Peptest and
Pepsin ELISA (P=0.55); Bland-Altman analysis suggested no agreement between the tests
(P=0.414) implying that they do not measure the same analyte. Receiver-operator curve
suggests that neither Peptest (P=0.3328) nor pepsin (P=0.4476) is useful for predicting
GERD .
Conclusions
Salivary pepsin is not a reliable tool for the diagnosis of GERD.
198 words
Page 2 of 24Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
INTRODUCTION
Gastro-esophageal reflux disease (GERD) is one of the most common gastrointestinal
disorders and is defined as a condition that develops when the reflux of stomach contents
causes troublesome symptoms and/or complications 1. Although the typical symptoms that
accompany reflux into the oesophagus are heartburn and acid regurgitation, non-specific
symptoms are not uncommon 1. Furthermore, when gastric refluxate passes the upper
esophageal sphincter (EOR) a range of laryngeal, pharyngeal and respiratory symptoms
including cough and recurrent infections may result 2.
Current diagnostic methods for GERD include symptomatic assessment (including
structured questionnaires 3, therapeutic trials with proton pump inhibitors 4,5, endoscopy 6
and ambulatory esophageal reflux monitoring 7). Unfortunately, all have limited sensitivity
and specificity 6 furthermore the latter two are invasive. There is therefore an unmet need for
non-invasive diagnostic tools with good sensitivity and specificity for this condition.
Pepsin is a protease which is synthesised, via its precursor pepsinogen, in gastric chief cells
8 and its presence in the oesophagus or more proximal sites is argued to be indicative of
reflux 9. Interestingly pepsin has been reported in laryngeal and nasal sinus tissues, tracheal
secretions and broncho-alveolar lavage fluid and in saliva 9.
Recent reports have suggested that salivary and/or sputum pepsin may be a clinically useful,
non-invasive diagnostic marker for GERD. A commercially developed lateral flow test device
for detecting and quantifying pepsin as a biomarker of reflux in saliva and sputum samples
has been released. Although no definitive evaluation is available in the peer-reviewed
literature, the five studies using it to date have yielded conflicting results 10-14 and NICE
remains cautious about endorsing its use. Peptest comes with little information on the
optimal timing for sampling, persistence of pepsin in the mouth following reflux, or indicative
limits of detection, or discriminatory potential from the normal population. The aim of this
study was to assess the best time to take samples for salivary pepsin assay and to validate
the findings against both conventional reflux testing and against a validatable pepsin assay
in both GERD patients and normal volunteers.
Page 3 of 24 Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
METHODS
Study Design
This was a case-control comparison of a commercial diagnostic assay. Cases were
symptomatic patients referred for anti-reflux surgery, controls were asymptomatic volunteers.
The study had two parts: (i) an initial pilot assessed feasibility of diurnal repeated saliva
sampling, collection and posting on the day of return by the patient, all following
manufacturer recommendations, in 10 cases and 10 controls; (ii) an extended study using
the same sampling and storage protocol in a further 20 cases and 10 controls, with samples
split and also analysed by a second independent method.
Subjects
Studies were performed on asymptomatic healthy volunteers and patients referred for
consideration of anti-reflux surgery. Healthy volunteers were recruited by advertisements
placed in Sheffield Teaching Hospital NHS Trusts. Volunteers with a history of upper
gastrointestinal symptoms or those taking medications known to influence upper
gastrointestinal function were excluded. Patients were recruited from referrals to the
Gastrointestinal Physiology Unit for esophageal reflux and motility testing prior to
consideration of a surgical antireflux procedure. All had typical reflux symptoms and had had
a good or partial response to proton pump inhibitor treatment. Patients known to have other
non-reflux related upper gastrointestinal disease or general disorders or drugs (other than
acid suppression) known to influence upper gastrointestinal function were excluded. The
study was approved by the Sheffield Teaching Hospital NHS Ethics Committee (NRES
Committee Yorkshire & Humber 12/ YH/0466) and Sheffield Research Department
(STH16373).
Study Protocol
Following written informed consent all subjects underwent a detailed symptom assessment
and drug history and were asked to complete two validated symptom questionnaires which
have a high sensitivity for reflux disease (the Gastro-esophageal Reflux Score Questionnaire
(GORS) 3 and the Hull Reflux Cough Questionnaire (HARQ) 15. Selection was based on
positive GORS (≥4) with HARQ (≥13) included for broader assessment of symptoms. Clinical
discussion occurred on the day to confirm the symptoms described by patients agreed with
scores before inclusion. All participants were studied off proton pump inhibitor therapy for at
least 7 days (pharmacokinetic analysis of PPI suggests that clearance occurs with a 60
minutes half-life, and furthermore that PPs have a half-life of 56hr and a daily turnover rate
of 20%, supporting 7 days as an adequate wash-out window16 in line with previous
Page 4 of 24Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
assessments of Peptest17), H2RAs for 2 days and over the counter antacids and alginates for
24 hours.
Subjects were asked to fast for 4 hours prior to study and then underwent high-resolution
esophageal manometry and 24-hour dual pH testing and produced saliva samples for pepsin
analysis as detailed below.
High resolution manometry was performed using Manoscan equipment (Sierra Scientific
Instruments, Los Angeles USA). Esophageal pH testing was performed using a dual channel
pH catheter (Versaflex®Given Imaging, Mansfield USA) attached to a novel data logger/
alarm device (PDTronics, Sheffield, UK). Since acid reflux episodes are often asymptomatic
we developed a device to fulfil the normal functions of a pH datalogger but also produce an
audible and visible alarm when the proximal pH channel fell below pH 4. This prompted the
subjects to start collecting saliva samples following the first detected proximal acid reflux
events. Following interim analysis of the 10 volunteers and 10 patients, esophageal
impedance was additionally performed using the Sandhill system (Sandhill Scientific
(Diversatek Healthcare, Milwalkee USA). Prior to the 24 hour study, pH catheters were
placed such that the sensors were 5 and 20 cm above the manometrically determined lower
esophageal sphincter (LOS) and the six impedance sensors were placed at 3, 5, 7, 9, 15
and 17 cm above the LOS. All subjects were asked to record symptoms, meal times and
periods of recumbency.
All tracings were reviewed manually to ensure accurate reflux detection. Proximal reflux was
defined as refluxate reaching the 20 cm pH and or 17-cm impedance sensor. Symptom
index (SI) and Symptom Association Probability (SAP) were used to characterise the
association between symptoms and reflux.
Saliva sample collection and analysis
In order to identify the optimal timing of saliva sampling for pepsin analysis all subjects were
asked to collect saliva samples; 2, 5, 10, 15, 30 and 60 minutes after reflux symptoms and
after activation of the proximal reflux alarm. Since reflux is common after meals and periods
of recumbency subjects were also asked to collect similarly timed saliva samples on rising in
the morning (before eating, drinking or teeth brushing) and after the evening meal. Saliva
was collected following RD Biomed sampling and packaging instructions and using their
buffer (0.5 ml of 0.01 M citric acid) and were stored as recommended (4°C) and returned at
the end of the 24-hour study period on the same working day (as per RD Biomed sampling
instructions). All samples were coded and analyses were performed blind to all clinical and
physiological variables. For the extended study all salivary samples were homogenised by
vortex mixer and split in two, one half sent for Peptest analysis (Peptest, RD Biomed, Hull),
the second half was quantified for pepsin by ELISA at The University of Sheffield. Samples
Page 5 of 24 Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
received on ice at The University of Sheffield were frozen immediately and stored at -80°C
prior to assay.
Peptest assay All salivary samples were sent to RD Biomed for proprietary analysis of
pepsin (Peptest, RD Biomed, Hull) which was undertaken as paid service provision (i.e. the
samples were presented as for diagnosis, not part of a research collaboration).
Quantification of pepsin by Peptest was provided by RD Biomed as a data sheet via email
usually with 1-2 working days of the sample being sent.
ELISA pepsin assay was an indirect ELISA using antiPepsin A monoclonal antibody (Santa
Cruz sc-365680, raised against an antigen of aa281-324 of Pepsin A). The ELISA was
calibrated using recombinant human pepsin A (Stratech SCB-RP112894h, which includes
aa15-388 of Pepsin A). The ELISA was linear when tested in the range from 0-150ng with a
limit of detection of 2ng. For development and validation of the ELISA see Supplementary
Online Information (SOI, Section 1).
Statistical analyses
Statistics were undertaken as indicated using SPSS 23 (IBM, Armonk, New York, 2015),
other than Bland and Altman Tests, which were produced using R (3.4.1), with the packages
BlandAltmanLeh (v 0.3.1) and ggplot2 (v2.2.1).
Page 6 of 24Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RESULTS
Subjects and demographics
Control group: 50 volunteers responded to the advert; 20 (6 male, 14 female, aged 21 to 56
years) met the inclusion criteria and completed the study, 3 had exclusion criteria, 26
declined intubation and one was unable to tolerate intubation
Patient group: 101 patients were approached; 30 (18 male, 12 female, aged 20 to 72 years)
met the inclusion criteria and completed the study, 49 declined to participate, 13 were
excluded as they did not meet the symptom criteria, 8 were excluded as the pH study was
negative at the time of processing the salivary samples and one was unable to tolerate the
intubations. All the volunteers denied reflux symptoms and all 30 patients reported regular
heartburn and or regurgitation. Questionnaire scores were significantly different between the
two groups (GORS: volunteers 0(0-0) versus patients 10(3-15) p < 0.005; HARQ: volunteers
0(0-9) versus patients 28 (1-65) (p< 0.005).
All 20 volunteers had esophageal acid exposure times within the normal range (median
1.0%, range 0 - 3.5%). The 10 volunteers who also underwent MII had values for non-acid
and gas reflux within the normal range (mean total number of reflux events 14.6). Of the 30
patients studied 23 had esophageal acid exposure times in excess of the normal range
(Median 8.85, and range 3.0 (one individual with significant supine reflux) – 39%) and were
classified as having true acid reflux. A further 3 patients had normal acid exposures (Median
2.85, range 2.6-4.6%) but had abnormal non-acidic reflux (Median 81, range 74-101 events).
Four patients were found to have normal reflux values on the day of study and all
demonstrated a negative SAP and were therefore deemed to have functional heartburn and
were excluded. A positive SAP was classified as >95%.
In patients with true acid reflux, 20 reported symptoms during the 24-hour period and in 12
the proximal reflux alarm was triggered. Manual analysis of the dual pH tracings confirmed
that such episodes were all associated with both distal and proximal esophageal
acidification.
Salivary Pepsin analysis
In the initial pilot study, the ten volunteers produced a total of 138 salivary samples for
Peptest analysis. Seven of the volunteers had detectable levels at reported concentrations of
25 – 250 ng/ml, often in multiple samples. Of the seven patients who had true reflux, all had
detectable levels at concentrations of 25 – 250 ng/ml in at least one and often in several
samples. The lack of a clear difference between asymptomatic volunteers and patients
Page 7 of 24 Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
prompted us to investigate the validity of the Peptest. We therefore extended the study to a
further 10 volunteers and 20 patients following an identical protocol but splitting saliva
samples for analysis by both Peptest and an in-house indirect ELISA.
Peptest analysis of 276 samples from the 20 volunteers revealed detectable levels in 102
samples in total (37%). 75 % of volunteers had at least one positive sample for salivary
pepsin (range 1-18 positive samples). Peptest analysis of 458 samples from 25 of the 30
patients with true reflux revealed positive results in 41 % of samples. 84 % of patients had
at least one positive result. The indirect ELISA of 141 samples from the 10 volunteers
revealed detectable levels in 122 samples (87 %). All 10 (100 %) of volunteers had at least
one positive sample for salivary pepsin. Analysis of 393 samples from 20 patients with
pathological reflux revealed positive results in 309 (79%) of samples. All 20 (100 %) of
patients had at least one positive result.
The mean scores and ranges between the asymptomatic control group and the GERD-
diagnosed patients were compared (Fig 1A). Peptest gave a significant difference in score
between the control and GERD groups, but the controls gave the higher reading (t-test,
p<0.01). In contrast, there was no significant difference in the average pepsin concentration
in each group by ELISA. There were significant differences between the readouts of the two
quantitation methods within groups (Mann-Whitney, p<0.01 for Control arm, p<0.001 for
GERD arm).
To assess whether timing of sampling influenced outcome the data were also analysed
according to sampling point (rising, post-prandial, post-alarm, post symptoms), and
differences between groups tested. Peptest quantitations (Fig 1B) were not significantly
different between patients and volunteers after rising or evening meal, but were significantly
lower in the patient group on reflux alarm (p<0.05, without Bonferroni correction). A
comparison was undertaken between Peptest scores for 2-60 minutes following reflux alarm
in controls and patients (Fig1C). The Peptest results showed no difference in salivary
readout across time, but symptomatic individuals had lower readouts than the control group
(P<0.05, one way ANOVA). We assessed the value Peptest as a diagnostic tool using a
receiver-operator (ROC) analysis (Figs 1D). Peptest yielded an area-under-the-curve of
0.528 with P=0.333 indicated the test is worthless.
Finally, a linear mixed model 18was deployed to assess contributions to variance allowing for
the repeated nature of measures with respect to time using a First Order autoregressive, Fig
1E. No variable explained significantly any aspect of the variation in data. In summary, the
Page 8 of 24Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
readouts for Peptest were lower for subjects with diagnosed GERD than for asymptomatic
controls when all samples were considered. Furthermore Peptest readouts were not
significantly different in either fasted state or following reflux for subjects with GERD than
asymptomatic volunteers.
Comparison of Peptest and ELISA analyses
As distinct outcomes were found for the Peptest and ELISA analyses (see section 2) direct
comparisons were undertaken between the quantitations from the two analyses.
Quantifications were compared in several ways (Fig 2). Quantitations from each method
were plotted against each other and their relationship tested using a Spearman’s correlation
undertaken (Fig 2A). This test did not indicate any statistically detectable relationship
between the outcomes from each test.
As the data did not correlate we assessed their distributions independently. The ELISA data
followed a biologically usual gamma-distribution (SOI section 2). The Peptest data did not
follow a gamma distribution and show large frequencies outside the measurement limits
(SOI, Section2): the data were in a range of 0-500ng, but 73% of the points were outside the
limits of the sensitivity of measure (63% below and 10% above). To try to identify
relationships, data were split into three categories according to Peptest result (0, 500 and
0<datapoint<500). The distributions of ELISA data for each Peptest category were analysed
and shown to have no significant difference (SOI, Section 2). Bland-Altman analysis is a
standard statistical option for comparing between two tests of the same endpoint, and was
used to assess whether the assays were measuring the same substance (Fig 2Bi and ii).
The plots shows a cluster of datapoints where both tests are at their minimal point. Fig2Bi
presents untransformed data which exhibit poor distribution due to the narrow range. Data
were therefore log transformed (Fig 2Bii). Whilst the distribution is wider, the paucity of data
in the sensitive range for Peptest distorts the graph. Intraclass Correlations (ICC) were
undertaken for both the untransformed and log transformed data. For the untransformed
data the ICC is 0.019 (p=0.414) for the transformed data the ICC is 0.54 (p=0.268). These
ICC data suggest that the two tests are not measuring the same endpoint. As the ELISA is
undertaken with a commercially validated antibody, revalidated against recombinant pepsin,
shown to be linear for pepsin, and follows a biologically common gamma distribution, we
deduce that it is reliably measuring pepsin. We therefore infer that Peptest is not measuring
pepsin.
Evaluation of salivary pepsin and Peptest as a diagnostic markers for GERD
The ELISA quantitations were then independently assessed in terms for potential to predict
GERD. When ELISA quantitations were assessed globally, there was no significant
Page 9 of 24 Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
difference between the amount of pepsin diagnosed and asymptomatic individuals (Fig 3A).
As for the Peptest analyses, quantitative data for pepsin were stratified by sampling time
(rising, postprandially, after alarm and after symptoms). There were no significant
differences in salivary pepsin concentration at any of these timepoints (Fig 3B). When pepsin
was analysed in the time period following a reflux alarm there were no significant differences
between volunteers and patients, and no significant changes across time (Fig 3C). Finally,
the data were dichotomised as RD Biomed have suggested a diagnostic threshold of
16ng/ml 17, which demonstrated rates of positives were identical in the GERD and control
groups (SOI, Section 3). Levels of pepsin in saliva did not discriminate between
asymptomatic subjects and patients diagnosed with GERD in any experimental setting.
Finally we assessed the value of the Pepsin ELISA as a diagnostic tool using ROC analysis
(Fig 3D). The area-under-the-curve analysis was 0.522 with P=0.448 indicating that the test
is worthless.
Page 10 of 24Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
DISCUSSION
Peptest is a highly promoted diagnostic device presented for assessment of reflux through
detection of pepsin. It utilizes a proprietary monoclonal antibody (only available to RD
Biomed) in a lateral flow format to quantify levels of pepsin in saliva (based on the working
hypothesis that pepsin in saliva is non-routine and can only be attributed to reflux). A recent
review of studies assessing salivary pepsin for LPR 19 erred to a conclusion that pepsin may
be of value, however studies were inconsistent with multiple methodologies deployed, a
range of cut-offs and differences in study design and in control populations. Only one study
using Peptest was of sufficient rigour for inclusion in that systematic review 12 reporting that
Peptest yielded the same rates of positives in cases and controls of LPR.
Our initial aim was to establish optimal timings for saliva sampling relative to diurnal variation
and to measure spiking and persistence of pepsin in saliva following reflux to support optimal
application of Peptest. However a very high rate of false positives (and negatives) relative to
clinical diagnostic criteria led to re-evaluation of both Peptest and pepsin data in the context
of GERD specifically. Our finding is that pepsin is present in control, asymptomatic subjects
and at measurable levels and does not have discriminatory potential for separating GERD
from controls. This finding is consistent with other reports which found no difference between
cases and controls 9,20,21, or found controls had higher levels 22. A recently published study
concluded a value for Peptest in predicting GERD, however there was no difference
between controls and GERD patients when all sampling was taking into account, suggesting
low reproducibility 10. A second recent paper assessing application of Peptest in GERD also
showed poor reproducibility of Peptest and depended on data-selection to achieve any
diagnostic power17. Peptest outcomes were provided to us by RD Biomed as a quantification
in ng/ml of pepsin in saliva. When analysed, the distribution of these data was very skewed
(Fig 2). As the analytical platform is a window in a lateral flow device, there may be very
significant potential for subjective score or operator error.
The presence of pepsin in saliva is thought to be indicative of gastric refluxate and is the
premise of test development, however recent data arising from the Fantom5 project 23
indicates expression of pepsin in the tongue (SOI, section 3). This outcome from a high-
throughput screen requires direct validation but provides a plausible explanation for
presence of pepsin in saliva in controls as an apparent inconsistency reported across
several studies.
A recent review and guideline advises the use of a constellation of scores in diagnosis of
GERD and shifts away from a single measure 24. Our report demonstrates inconsistencies
Page 11 of 24 Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
between ELISA quantitations (using a validated commercial antipepsin antibody, re-validated
and calibrated against pepsin from a different supplier) and Peptest quantitations, and
demonstrated an unusual distribution of data for Peptest by two separate statistical tests.
Further recent work questions the diagnostic value of Peptest / pepsin25. As far as we can
determine using a formal statistical test, Peptest and our pepsin ELISA are measuring
differing analytes. We suggest that Peptest is neither useful as a diagnostic for GERD, nor is
a measure of pepsin.
CONCLUSIONS
There is an unmet need for non-invasive diagnostics for GERD, salivary pepsin has
been proposed as one such test;
Our data and those of others indicate pepsin does not discriminate effectively
between healthy asymptomatic controls and patients with confirmed reflux;
Our data suggest that Peptest does not reliably measure pepsin.
ACKNOWLEDGEMENTS
CONFLICT OF INTEREST
Guarantor of the article: Dr Stuart Riley
Specific author contributions: SAR conceived the study, directed the clinical aspects,
oversaw data analysis and interpretation and revised the drafts; CR undertook the
recruitment, sampling and clinical assessments, managed the data and undertook
preliminary analysis, contributed to the drafting of the paper; JC undertook ELISA
development and validation, quantifications, undertook preliminary analysis and contributed
to the drafting of the paper; JMR undertook and directed statistical analyses and
interpretation; BMC contributed to experimental design and interpretation, directed the lab
analyses and drafted the paper. All authors approved the final submitted version of the
paper.
Financial support: none declared
Potential competing interest: none declared.
Page 12 of 24Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
REFERENCES
1. Kellerman R, Kintanar T. Gastroesophageal Reflux Disease. Primary Care. 2017;44(4):561-+.
2. Rafferty G, Mainie I, McGarvey LPA. Respiratory and laryngeal symptoms secondary to
gastro-oesophageal reflux. Frontline gastroenterology. 2011;2(4):212-217.
3. Carlsson R, Dent J, Bolling-Sternevald E, et al. The usefulness of a structured questionnaire in
the assessment of symptomatic gastroesophageal reflux disease. Scandinavian Journal of
Gastroenterology. 1998;33(10):1023-1029.
4. Bashashati M, Hejazi RA, Andrews CN, Storr MA. Gastroesophageal reflux symptoms not
responding to proton pump inhibitor: GERD, NERD, NARD, esophageal hypersensitivity or
dyspepsia? Canadian Journal of Gastroenterology and Hepatology. 2014;28(6):335-341.
5. Katz PO, Gerson LB, Vela MF. Guidelines for the Diagnosis and Management of
Gastroesophageal Reflux Disease (vol 108, pg 308, 2013). American Journal of
Gastroenterology. 2013;108(10):1672-1672.
6. Kahrilas P, Yadlapati R, Roman S. Emerging dilemmas in the diagnosis and management of
gastroesophageal reflux disease. F1000Research. 2017;6:1748-1748.
7. Roman S, Gyawali CP, Savarino E, et al. Ambulatory reflux monitoring for diagnosis of gastro-
esophageal reflux disease: Update of the Porto consensus and recommendations from an
international consensus group. Neurogastroenterology and Motility. 2017;29(10).
8. Muller MJ, Defize J, Hunt RH. CONTROL OF PEPSINOGEN SYNTHESIS AND SECRETION.
Gastroenterology Clinics of North America. 1990;19(1):27-40.
9. Kim TH, Lee KJ, Yeo M, Kim DK, Cho SW. Pepsin detection in the sputum/saliva for the
diagnosis of gastroesophageal reflux disease in patients with clinically suspected atypical
gastroesophageal reflux disease symptoms. Digestion. 2008;77(3-4):201-206.
10. Du X, Wang F, Hu Z, et al. The diagnostic value of pepsin detection in saliva for gastro-
esophageal reflux disease: a preliminary study from China. Bmc Gastroenterology. 2017;17.
11. Dy F, Amirault J, Mitchell PD, Rosen R. Salivary Pepsin Lacks Sensitivity as a Diagnostic Tool
to Evaluate Extraesophageal Reflux Disease. Journal of Pediatrics. 2016;177:53-58.
12. Yadlapati R, Adkins C, Jaiyeola D-M, et al. Abilities of Oropharyngeal pH Tests and Salivary
Pepsin Analysis to Discriminate Between Asymptomatic Volunteers and Subjects With
Symptoms of Laryngeal Irritation. Clinical Gastroenterology and Hepatology. 2016;14(4):535-
+.
13. Formanek M, Kominek P, Matousek P, Tomanova R, Urban O, Zelenik K. Comparison of
Three Methods Used in the Diagnosis of Extraesophageal Reflux in Children with Chronic
Otitis Media with Effusion. Gastroenterology Research and Practice. 2015.
14. Formanek M, Zelenik K, Kominek P, Matousek P. Diagnosis of extraesophageal reflux in
children with chronic otitis media with effusion using Peptest. International Journal of
Pediatric Otorhinolaryngology. 2015;79(5):677-679.
15. Morice AH, Faruqi S, Wright CE, Thompson R, Bland JM. Cough Hypersensitivity Syndrome: A
Distinct Clinical Entity. Lung. 2011;189(1):73-79.
16. Shin JM, Kim N. Pharmacokinetics and Pharmacodynamics of the Proton Pump Inhibitors. J
Neurogastroenterol Motil. 2013;19(1):25-35.
17. Hayat JO, Gabieta-Somnez S, Yazaki E, et al. Pepsin in saliva for the diagnosis of gastro-
oesophageal reflux disease. Gut. 2015;64(3):373-380.
18. Rodrigues AM. Silva Lusitana. 2009;17(1):123-125.
19. Calvo-Henriquez C, Ruano-Ravina A, Vaamonde P, Martinez-Capoccioni G, Martin-Martin C.
Is Pepsin a Reliable Marker of Laryngopharyngeal Reflux? A Systematic Review.
Otolaryngology-Head and Neck Surgery. 2017;157(3):385-391.
20. Iannella G, Di Nardo G, Plateroti R, et al. Investigation of pepsin in tears of children with
laryngopharyngeal reflux disease. International Journal of Pediatric Otorhinolaryngology.
2015;79(12):2312-2315.
Page 13 of 24 Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
21. Birtic D, Vceva A, Kotromanovic Z, Zubcic Z, Mihalj H, Jovanovic S. Significance of the Pepsin
from the Saliva in the Diagnosis and Treatment of Laryngopharyngeal Reflux Disease.
Collegium Antropologicum. 2012;36:83-86.
22. Komatsu Y, Kelly LA, Zaidi AH, et al. Hypopharyngeal pepsin and Sep70 as diagnostic markers
of laryngopharyngeal reflux: preliminary study. Surgical Endoscopy and Other Interventional
Techniques. 2015;29(5):1080-1087.
23. Andersson R, Gebhard C, Miguel-Escalada I, et al. An atlas of active enhancers across human
cell types and tissues. Nature. 2014;507(7493):455-+.
24. Vaezi MF, Katzka D, Zerbib F. Extraesophageal Symptoms and Diseases Attributed to GERD:
Where is the Pendulum Swinging Now? Clinical gastroenterology and hepatology : the
official clinical practice journal of the American Gastroenterological Association. 2018.
25. Woodland P, Singendonk MMJ, Ooi J, et al. Measurement of Salivary Pepsin to Detect
Gastroesophageal Reflux Disease Is Not Ready for Clinical Application. Clinical
Gastroenterology and Hepatology.in press
Page 14 of 24Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
LEGENDS
Figure 1. Analysis of discrimination between patients and controls by Peptest.
Panel A shows a comparison of Peptest quantitations for all samples for the controls (open
circles) and GERD patients (closed circles). The average in controls was significantly higher
than patients (p<0.005, unpaired t-test). Panel B compares the mean readout for Peptest,
stratified by sampling time and by patients (grey filled bars) or controls (white bars). There
was no significant difference between patients and controls at any time except following the
reflux alarm, when controls were higher than patients. No controls reported a reflux
symptom, so no comparison was undertaken. Panel C shows Peptest readout across the 60
minutes following reflux alarm for controls (open circles) and patients (filled circles). There
was no significant difference across time. Patients scores were significantly and consistently
lower than controls (P,0.05, 1-way ANOVA). Event rate was very low in the control group
with no self-reported symptoms and only three alarms across the study. Panel D Receiver-
Operator Curve (ROC) analyses of Peptest data. The dataset had an AUC of 0.53
(P=0.333). Panel E shows variance component analysis using multi-level modelling
(Repeated Measures ANOVA) for GERD patients and with regards to event type and sample
time using a First Order autoregressive (SPSS 24). The results demonstrate no significance
except when sample taken during a reflux event.
Figure 2. Comparison of Peptest and pepsin ELISA.
Panel A shows Spearman’s correlation between quantifications of pepsin in Saliva from both
Peptest (x-axis) and indirect ELISA (y-axis). Correlation yielded a very weak r2, of 0.0007,
which was not significant. Panel B shows Bland-Altman plots for the same comparison:
Panel Bi is untransformed data (Intraclass correlation (ICC)=0.019, P=0.414), Panel Bii is
log-transformed data (ICC=0.54, P=0.268).
Figure 3. Pepsin does not predict GERD
Panel A shows a comparison of ELISA pepsin quantitations for all samples for the controls
(open circles) and GERD patients (closed circles). There was no difference between groups
(unpaired t-test). Note the difference in dynamic range of these data (0-100) by comparison
with Peptest scores (Fig1A, range 0-600)
Panel B compares the mean readout for pepsin, stratified by sampling time and by patients
(grey filled bars) or controls (white bars). There was no significant difference between
patients and controls at any time. No controls reported a reflux symptom, so no comparison
was undertaken. Panel C shows pepsin ELISA readout across the 60 minutes following
reflux alarm for controls (open circles) and patients (filled circles). There was no significant
Page 15 of 24 Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
difference across time (1-way ANOVA). Panel D Receiver-Operator Curve (ROC) analyses
of pepsin ELISA data. The dataset had an AUC of 0.52 (P=0.448).
Page 16 of 24Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
A
E
C
B
Source NumeratorDf
Denominatordf
F Sig.
EventType 3 51.91 0.424
0.736
SampleTime 5 208.13 2.668
0.023
EventType*SampleTime 15 207.95 1.184
0.286
GORDDiagnosis 1 34.74 0.036
0.851
EventType*GORDDiagnosis 3 51.91 0.324
0.808
SampleTime*GORDDiagnosis 5 208.131 0.466
0.801
EventType*SampleTime*GORDDiagnosis
15 207.95 1.046
0.410
Control GORD
Salivarypepsin(ng
/ml)
AssayedbyPeptest
Salivarypepsin(ng
/ml)
AssayedbyPeptest
D
Page 17 of 24 Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
For Peer Review
R2=0.0007P=0.55Y=0.0025x+13.05
ICC=0.019P=0.414
ICC=0.54P=0.268
A
Bi
ii
Page 18 of 24Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
For Peer Review
A
C
B
Control GORD
Salivarypepsin(ng
/ml)
AssayedbyEL
ISA
Salivarypepsin(ng
/ml)
AssayedbyEL
ISA
D
Page 19 of 24 Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
For Peer Review
Supplementary Online File
for:
Studies of Salivary Pepsin in Patients with GERD
Caroline Race1,5 Joanna Chowdry2,5 Jean M. Russell3, Bernard M. Corfe2,4,6 & Stuart
A. Riley1,4,6,7
Page 20 of 24Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
Section 1 – Pepsin ELISA Development
Determination of specificity of anti-pepsin monoclonal antibody.
Available monoclonal antibodies were surveyed for those with good user ratings, and proven application
in both western immunoblot and ELISA. Santa Cruz sc-365680 was selected. The sensitivity of the
antibody was determined in both immunoblot and ELISA.
Left Panel For immunoblot assessment recombinant human pepsin A was loaded at 0-100ng into SDS-
PAGE. The antibody cross-reacted with a single band (although breakdown products were detected at
high antigen concentrations). The LoD for immunoblotting appeared to be <10ng.
Right Panel For ELISA, an indirect ELISA was developed and plates were coated with 0-40ng of
recombinant pepsin, cross reactions were visualised with horseradish peroxidase conjugated anti-mouse
secondary antibody. A linear relationship was observed in the 1-40ng range.
Page 21 of 24 Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
Section 2 – Comparison of Peptest and ELISA results
The distributions of data from 499 salivary analyses for pepsin ELISA and for Peptest are shown below.
Peptest returned a large number of 0 scores which were converted to 8 (half of the stated limit of
detection) in order to process the data. Pepsin ELISA follows a classical gamma distribution, whereas the
Peptest data is hyperskewed with 73% of datapoints at either minimus or maximus. Panel C shows the
distribution of pepsin quantifications according to three strata of Peptest scores (Upper boundary, lower
boundary and sensitive range). There is no evidence of a significant difference in the distribution of
pepsin ELISA data in these three categories (Kruskall Wallis=0.627. p=0.731).
Pepsin level (ng/ml) from ELISA (zeros and below coded to 1)
Page 22 of 24Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
Section 3 – Dichotomous analyses of Peptest and ELISA results according
to RD Biomed’s suggested thresholds
The data from Peptest and ELISA scores were dichotomised at 16ng/ml (RD Biomed’s suggested
diagnostic threshold) and were compared (Panel A). The rates of postives in the Control and GORD
group are the same using this threshold by both methods. The Peptest data are hyperskwed: when only
values of 0 were considered (Panel B), there was no difference in the negative score rate between GORD
patients and controls. When only values att he upper limit of quantification were considered (Panel C)
there are more positives in the control group.
Page 23 of 24 Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
Section 3 – Expression of pepsinogen in tongue
Expression analysis was undertaken for the expression profile of pepsin (the precursor peptide
pepsinogen is shown). Analysis shown below indicates expression includes the tongue, suggesting
potential for oral pepsin antigens. Expression Atlas accessed on 21st March 2018.
Page 24 of 24Alimentary Pharmacology & Therapeutic
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60